Linear Algebra

Lecture 02

Prof. Shreedevi K. Masuti

IIT Dharwad

06 Jan 2020

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Outline

• Summary of Lecture 01
• Matrix multiplication revisited

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Matrix Multiplication
 
Let n ∈ N, a := a1 · · · an ∈ M1,n be a row vector and
 
b1
 .. 
b :=  .  ∈ Mn,1 be a column vector.
bn
Step 1: Define the product of a row vector a with a column vector b as:
 
b1
  . 
ab = a1 · · · an  ..  := a1 b1 + · · · + an bn ∈ R.
bn

(This is like the scalar product of x and y in Rn .)

Step 2: Let m ∈ N and A ∈ Mm,n and let R1 , R2 , . . . , Rm be the rows of
A. Then we define
   
R1 R1 b
Ab =  ... b :=  ...  ∈ Mm,1 .
   

Rm Rm b

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

Step 3: Finally, let p ∈ N and B ∈ Mn,p . Then B = [C1 · · · Cp ], where
C1 , . . . , Cp are the columns of B. Then
   
R1 R1 C1 · · · R1 Cp
AB =  ...  [C1 · · · Cp ] =  ... ..  ∈ M .
  
.  m,p
Rm Rm C1 · · · Rm Cp

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Precise Forula

More precisely, if m, n, p ∈ N, A := [aij ] ∈ Mm,n and B := [bij ] ∈ Mn,p ,

then AB ∈ Mm,p , and for i = 1, . . . , m; j = 1 . . . , p,
Pn
AB = [cij ], where cij := ai bj = k=1 aik bkj

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

Example Let A be the matrix representing Laura’s timetable:
L1=Lectures (each 1.5 hours), T=Tutorials (1 hour), L2=Lab (3 hours)
Day/Classes L1 T L2
M 1 1 0
T 2 0 1
A= .
W 1 1 1
T 2 1 0
F 1 1 0
 t
Let B = 1.5 1 3 .
Question: How much hours does Laura has classes each day ?
This is given by the matrix AB

Day/Hours Class Hours

M 2.5
T 6
AB = .
W 5.5
T 4
F 2.5

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

Note:(1) The product AB is defined only when the number of columns of
A is equal to the number of rows of B.
Verify: AI = A, IA = A, AO = O and AO = O whenever these products
are defined

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Mysterious

Caution: (1) Multiplication is not commutative, that is, in general,

AB 6= BA.
(2) Cancellation law does not hold.

Example
       
1 0 0 1 0 1 0 0
A := and B := , then AB := , while BA :=
0 0 0 0 0 0 0 0
Moreover, AB = B, but A 6= I .

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Important Remark

Proposition
Let A := [aij ] ∈ Mm,n , and let e1 , . . . , en be the basic column vectors in
Mn,1 . Then for k = 1, . . . , n,
 
a1k
 .. 
 . 
 
A ek =  ajk , which is the kth column of A.

 .. 
 . 
amk

Corollary
A = B if and only if Aek = Bek for each k = 1, . . . , n.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Outer Product
 
c1
 .. 
 . 
 
 cj  ∈ Mm,1 be a column vector of length m, and let
Let c :=  
 .. 
 . 
 cm 
d := d1 · · · dk · · · dp ∈ M1,p be a row vector of length p. Then
 
c 1 d1 ··· c1 dj ··· c1 dp
 .. .. .. .. .. 
 . . . . . 
 
 ci d1
cd =  ··· ci dj ··· c i dp  = [ci dj ] ∈ Mm,p .
 .. .. .. .. .. 
 . . . . . 
cm d1 ··· cm dj ··· c m dp

This is called the outer product of the column vector c and the row
vector d.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Matrix Multiplication revisited

What if A is written in terms of column vectors and B in terms of row

vectors ?
 
R1
   .. 
Let A = 1 C · · · C n and B =  . ,
Rn
Thus
 
R1
Cn  ...  = C1 R1 + · · · + Cn Rn ∈ Mm,p .
  
AB = C1 ···
Rn

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Properties of Matrix Multiplication

Proposition
For the matrices A, B, C (of appropriate order so that the sums &
products are well-defined) and α ∈ R,
(1) Distributive law: (A + B)C = AC + BC , C (A + B) = CA + CB and
(2) (αA)B = αAB = A(αB).

Proof: Exercise
Matrix multiplication also satisfies the associative law:
Proposition
Let m, n, p, q ∈ N, and A ∈ Mm,n , B ∈ Mn,p and C ∈ Mp,q . Then
A(BC ) = (AB)C (which we shall write as ABC ).

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

Also, the transpose of a product is the product of the transposes in the
reverse order:
Proposition
Let m, n, p ∈ N. If A ∈ Mm,n and B ∈ Mn,p , then (AB)t = B t At .

Proof: Let A := [aij ], B := [bij ] and AB := [cij ].

Also, let At := [aij0 ], B t := [bij0 ] and (AB)t := [cij0 ]. Then
n
X n
X
cij = aik bkj and so cij0 = cji = ajk bki
k=1 k=1

for i = 1, . . . , m; j = 1, . . . , p.
Suppose B t At := [dij ]. Then
n
X n
X
0 0
dij = bik akj = bki ajk = cij0
k=1 k=1

for i = 1, . . . , m; j = 1, . . . , p. Hence the result.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Linear Combination of Vectors

Definition
Let x1 , . . . , xk be vectors in Rn and α1 , . . . , αm ∈ R. Then

α1 x1 + · · · + αk xk

is called a (finite) linear combination of x1 , . . . , xk .

Example
 T
For k = 1, . . . , n, let ek := 0 · · · 1 · · · 0 ∈ Mn×1 be the
column vector, where the kth entry is 1 and all other entries are 0.
 t
If b = b1 · · · bk · · · bn is any column vector of length n, then it
follows that b = b1 e1 + · · · + bk ek + · · · + bn en , which is a linear
combination of e1 , . . . , en . The vectors e1 , . . . , en are known as the basic
column vectors.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

Proposition
Let A ∈ Mm,n and B ∈ Mn,p . Then (1) The ith row of AB is a linear
combination of the n row vectors of B with coefficients ai1 , . . . , ain
provided by the ith row of A
(2) The jth column of AB is a linear combination of the n column vectors
of A with coefficients b1j , . . . , bnj provided by the jth column of B

Proof: (1) Let AB = [cij ] and Rk be the k-th row of B. Then i-th row of
AB is
 
ci1 ci2 . . . cip
Pn Pn Pn 
= k=1 aik bk1 k=1 aik bk2 . . . k=1 aik bkp
   
= ai1 b11 b12 . . . b1p + ai2 b21 b22 . . . b2p + · · · +
 
ain bn1 bn2 . . . bnp
= ai1 R1 + ai2 R2 + · · · + ain Rn

(2) Proof is similar to (1).

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Example

Example
 
  1 6 0 2
2 1 −1
Let A = and B = 2 −1 1 −2. Then
0 3 1
2 0 −1 1
 
2 11 2 1
AB = .
8 −3 2 −5

We have
     
2 11 2 1 = 2 1 6 0 2 + 1 2 −1 1 −2
 
−1 2 0 −1 1 ,
       
11 2 1 −1
= 6 −1 +0 , etc.
−3 0 3 1
.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02

 Inverse of a matrix

Definition
Let A be an n × n matrix. If there is an n × n matrix B such that
AB = In = BA then we say A is invertible and B is the inverse of A. The
inverse of A is denoted by A−1 .

Remark: (1) If A is invertible, then it has a unique inverse.

(2) If A and B are invertible n × n matrices, then AB is also invertible.

(3) Inverse of a square matrix need not exist.

Proposition
For any invertible square matrix A, (At )−1 = (A−1 )t .

In particular, if A is symmetric, then so does A−1 .

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 02